stock-analyzer-skill 1.1.0

package/scripts/classifier.py

@@ -0,0 +1,302 @@
+#!/usr/bin/env python3
+"""
+A 股个股类型分类器。
+根据财务数据、行情数据和 K 线走势特征，自动判定 7 种类型。
+可被 technical.py 或其他模块 import 使用。
+"""
+from common import to_float, clamp
+
+
+# ── 类型 → 推荐指标映射 ──
+
+TYPE_INDICATOR_MAP = {
+    "题材股": {
+        "priority": ["K线形态", "涨停分析", "量比", "RSI"],
+        "deprioritized": ["MACD", "KDJ"],
+    },
+    "蓝筹股": {
+        "priority": ["均线系统", "BOLL", "MACD背离", "箱体"],
+        "deprioritized": ["KDJ"],
+    },
+    "强成长股": {
+        "priority": ["MACD趋势", "BOLL上轨", "量比", "RSI"],
+        "deprioritized": ["KDJ"],
+    },
+    "周期股": {
+        "priority": ["MACD背离", "KDJ钝化", "缠论背驰", "BOLL"],
+        "deprioritized": ["简单均线"],
+    },
+    "稳成长股": {
+        "priority": ["均线排列", "MACD", "BOLL", "RSI", "老鸭头"],
+        "deprioritized": ["KDJ"],
+    },
+    "防御股": {
+        "priority": ["BOLL", "RSI", "均线粘合", "箱体"],
+        "deprioritized": ["KDJ"],
+    },
+    "普通股": {
+        "priority": ["均线", "MACD", "KDJ", "BOLL", "RSI", "量价"],
+        "deprioritized": [],
+    },
+}
+
+
+def classify_stock(fin_record=None, quote_record=None, kline_records=None):
+    """
+    判定 A 股个股类型。
+
+    Args:
+        fin_record: 单期财务数据 dict（finance.py 返回的 records[0]）
+                   字段: ROEJQ, PARENTNETPROFITTZ, XSMLL, ZCFZL 等
+        quote_record: 实时行情 dict（quote.py 返回）
+                     字段: circulating_cap, total_cap, turnover, pe 等
+        kline_records: K 线数据 list
+
+    Returns:
+        {
+            "type": "强成长股"|...|"普通股",
+            "confidence": "高"|"中"|"低",
+            "reasons": [...],
+            "priority_indicators": [...],
+            "deprioritized": [...],
+        }
+    """
+    reasons = []
+    confidence = "高"
+
+    # 提取行情特征
+    circulating_cap = to_float(quote_record.get("circulating_cap")) if quote_record else 0
+    total_cap = to_float(quote_record.get("total_cap")) if quote_record else 0
+    turnover = to_float(quote_record.get("turnover")) if quote_record else 0
+
+    # 提取财务特征
+    has_finance = fin_record and isinstance(fin_record, dict)
+    roe = to_float(fin_record.get("ROEJQ")) if has_finance else 0
+    profit_growth = to_float(fin_record.get("PARENTNETPROFITTZ")) if has_finance else 0
+    gross_margin = to_float(fin_record.get("XSMLL")) if has_finance else 0
+    debt = to_float(fin_record.get("ZCFZL")) if has_finance else 0
+
+    # 连板检测
+    has_limit_streak = False
+    if kline_records and len(kline_records) >= 10:
+        from common import board_type as _board_type
+        code = quote_record.get("code", "") if quote_record else ""
+        bd = _board_type(code)
+        limit_ratio = {"主板": 9.5, "创业板": 19.5, "科创板": 19.5, "北交所": 29.5}.get(bd, 9.5)
+        for i in range(len(kline_records) - 1, max(len(kline_records) - 5, 0), -1):
+            r = kline_records[i]
+            prev = kline_records[i - 1] if i > 0 else r
+            chg = (to_float(r.get("close")) - to_float(prev.get("close"))) / max(to_float(prev.get("close")), 0.01) * 100
+            if chg >= limit_ratio * 0.95:
+                has_limit_streak = True
+                break
+
+    # ── 规则分类（优先级从高到低） ──
+    stock_type = "普通股"
+    confidence = "低"
+
+    # 1. 题材股
+    if circulating_cap > 0 and circulating_cap < 100 and (has_limit_streak or turnover > 10):
+        stock_type = "题材股"
+        reasons.append(f"流通市值{circulating_cap:.0f}亿(小盘)")
+        if has_limit_streak:
+            reasons.append("近期有连板记录")
+        if turnover > 10:
+            reasons.append(f"换手率{turnover:.1f}%(高活跃度)")
+        confidence = "高" if has_limit_streak else "中"
+
+    # 2. 蓝筹股
+    elif circulating_cap > 1000 and (not has_finance or roe > 10):
+        stock_type = "蓝筹股"
+        reasons.append(f"流通市值{circulating_cap:.0f}亿(大盘)")
+        if roe > 10:
+            reasons.append(f"ROE {roe:.1f}%")
+            confidence = "高"
+        else:
+            confidence = "中"
+
+    # 有财务数据时的精确分类
+    elif has_finance:
+        # 3. 强成长股
+        if profit_growth > 30 and roe > 15 and circulating_cap < 500:
+            stock_type = "强成长股"
+            reasons.append(f"净利增速{profit_growth:.0f}%")
+            reasons.append(f"ROE {roe:.1f}%")
+            reasons.append(f"流通市值{circulating_cap:.0f}亿")
+            confidence = "高"
+
+        # 4. 周期股（增速波动大+毛利偏低）
+        elif abs(profit_growth) > 50 and gross_margin > 0 and gross_margin < 25:
+            stock_type = "周期股"
+            reasons.append(f"净利增速波动大({profit_growth:.0f}%)")
+            reasons.append(f"毛利率{gross_margin:.1f}%(偏低)")
+            confidence = "中"
+
+        # 5. 稳成长股
+        elif 15 <= profit_growth <= 30 and roe > 12:
+            stock_type = "稳成长股"
+            reasons.append(f"净利增速{profit_growth:.0f}%")
+            reasons.append(f"ROE {roe:.1f}%")
+            confidence = "高"
+
+        # 6. 防御股
+        elif abs(profit_growth) < 20 and debt > 0 and debt < 50:
+            stock_type = "防御股"
+            reasons.append(f"增速稳定({profit_growth:.0f}%)")
+            reasons.append(f"负债率{debt:.1f}%(低)")
+            confidence = "中"
+
+        else:
+            stock_type = "普通股"
+            reasons.append("不满足以上特定类型条件")
+            confidence = "低"
+    else:
+        # 无财务数据，退化为市值分类
+        if circulating_cap > 500:
+            stock_type = "蓝筹股"
+            reasons.append(f"流通市值{circulating_cap:.0f}亿(推测大盘)")
+            confidence = "低"
+        elif circulating_cap < 100 and turnover > 5:
+            stock_type = "题材股"
+            reasons.append(f"流通市值{circulating_cap:.0f}亿(小盘)+高换手")
+            confidence = "低"
+        else:
+            stock_type = "普通股"
+            reasons.append("无财务数据，无法精确分类")
+            confidence = "低"
+
+    # ── 返回指标建议 ──
+    indicator_map = TYPE_INDICATOR_MAP.get(stock_type, TYPE_INDICATOR_MAP["普通股"])
+
+    return {
+        "type": stock_type,
+        "confidence": confidence,
+        "reasons": reasons,
+        "priority_indicators": indicator_map["priority"],
+        "deprioritized": indicator_map["deprioritized"],
+    }
+
+
+# ── 行业推断 ──
+
+def infer_industry(name: str, code: str = "") -> str:
+    """根据股票名称和代码推断行业分类（关键词匹配 + 代码段推断）。"""
+    name = name.upper()
+    # 科创板（688开头）默认归类为科技
+    if code.startswith("688") or code.startswith("sh688"):
+        return "科技"
+    # 金融子类：银行、券商
+    if any(kw in name for kw in ["银行"]):
+        return "银行"
+    if any(kw in name for kw in ["证券", "券商"]):
+        return "券商"
+    # 金融（保险、信托等）
+    if any(kw in name for kw in ["保险", "信托", "金融", "资管", "期货"]):
+        return "金融"
+    # 地产
+    if any(kw in name for kw in ["地产", "置业", "置地", "房产", "万科", "保利",
+                                  "碧桂园", "融创", "恒大", "绿地", "华润置地"]):
+        return "地产"
+    # 医药
+    if any(kw in name for kw in ["医药", "药业", "制药", "生物", "疫苗", "医疗",
+                                  "器械", "基因", "诊断", "CRO", "创新药"]):
+        return "医药"
+    # 科技子类：半导体、软件
+    if any(kw in name for kw in ["半导体", "芯片", "晶圆", "封测", "存储", "EDA"]):
+        return "半导体"
+    if any(kw in name for kw in ["软件", "信息", "数据", "云计算", "人工智能",
+                                  "物联网", "网络", "SaaS", "数字"]):
+        return "软件"
+    # 科技（其他）
+    if any(kw in name for kw in ["科技", "智能", "电子", "通信", "计算",
+                                  "技术", "威视", "光电", "机器人",
+                                  "安防", "集成", "激光", "光学", "显示", "面板"]):
+        return "科技"
+    # 消费
+    if any(kw in name for kw in ["白酒", "食品", "饮料", "乳业", "调味", "啤酒",
+                                  "茅台", "五粮液", "海天", "伊利", "牧原",
+                                  "家电", "美的", "格力", "海尔"]):
+        return "消费"
+    # 能源
+    if any(kw in name for kw in ["石油", "煤炭", "天然气", "能源", "石化", "燃气",
+                                  "中石油", "中石化", "中海油"]):
+        return "能源"
+    # 周期
+    if any(kw in name for kw in ["钢铁", "有色", "铜", "铝", "锌", "黄金", "矿业",
+                                  "化工", "化纤", "水泥", "稀土", "锂"]):
+        return "周期"
+    # 制造
+    if any(kw in name for kw in ["汽车", "机械", "制造", "装备", "新能源", "电池",
+                                  "光伏", "风电", "宁德", "比亚迪"]):
+        return "制造"
+    # 军工
+    if any(kw in name for kw in ["军工", "航天", "航空", "兵器", "船舶", "电科"]):
+        return "军工"
+    # 交运
+    if any(kw in name for kw in ["航空", "航运", "港口", "铁路", "高速", "物流"]):
+        return "交运"
+    return "默认"
+
+
+# ── 统一画像入口 ──
+
+def profile_stock(quote: dict, fin: dict = None, kline_records: list = None) -> dict:
+    """统一个股画像：行业 + 类型 + 指标建议。
+
+    Args:
+        quote: 行情 dict（含 code, name 等）
+        fin: 财务数据 dict（可选）
+        kline_records: K 线数据列表（可选）
+
+    Returns:
+        {
+            "industry": "金融"|...|"默认",
+            "type": "蓝筹股"|...|"普通股",
+            "confidence": "高"|"中"|"低",
+            "reasons": [...],
+            "priority_indicators": [...],
+            "deprioritized": [...],
+        }
+    """
+    name = quote.get("name", "")
+    code = quote.get("code", "")
+
+    industry = infer_industry(name, code)
+    classification = classify_stock(fin, quote, kline_records)
+
+    return {
+        "industry": industry,
+        **classification,
+    }
+
+
+# ── 命令行快速测试 ──
+if __name__ == "__main__":
+    import sys
+    import json
+
+    if len(sys.argv) < 2:
+        print("用法: python3 classifier.py <code>")
+        sys.exit(1)
+
+    from common import normalize_quote_code, normalize_finance_code
+    from quote import fetch_batch
+    from kline import fetch as fetch_kline
+
+    code = normalize_quote_code(sys.argv[1])
+    quotes = fetch_batch([code])
+    quote = quotes[0] if quotes else {}
+    records = fetch_kline(code, 240, 100)
+
+    # 尝试拉取财务数据
+    fin_record = None
+    try:
+        from finance import fetch as fetch_finance
+        fn_code = normalize_finance_code(code)
+        fin_data = fetch_finance(fn_code)
+        fin_record = fin_data[0] if fin_data else None
+    except Exception:
+        pass
+
+    result = classify_stock(fin_record, quote, records)
+    print(json.dumps(result, ensure_ascii=False, indent=2))